Page - (000020) - in Knowledge and Networks

8 depend on a variety of factors. The effect of the centrality of firms in the interorga- nizational network was very much a function of proximity: Highly central firms were more likely to be innovative if sited close to other firms than if sited far away from them. This finding corresponds with those reported in more recent studies in the context of trade fairs (e.g., Brailly, Favre, Chatellet, & Lazega, 2016a; Favre, Brailly, Chatellet, & Lazega, 2016; Piña-Stranger & Lazega, 2010). Connectivity as a Moderator of Spatial Effects on Knowledge The same kind of effect seems to be at work in the opposite case. It is usually accepted that information transfer and knowledge spillovers dwindle with geo- graphical distance. In the context of international technology transfer between units of multinational corporations, Hansen and Løvås (2004) explicitly focused on inter- action effects between the major factors of technology transfer. Their analysis con- veyed that the relations between distributed organizational units clearly moderate the association between technology transfer and geography. Units were found more likely to transfer technology successfully over large distances if they were con- nected through interpersonal ties or through formal organizational linkages than if there were no such links. Bell and Zaheer (2007) suggested that the kind of relation- ship, whether at the individual, organizational, or institutional level, varies in its dependence on geographical proximity. Empirically, they reported rather counterin- tuitive evidence for what they call geographic holes, that is, situations where infor- mation flow is facilitated between friends when they are geographically distant. In a similar vein, research on contractual alliances (Rosenkopf & Almeida, 2003) and informal business relationships alike (Glückler, 2006) illustrates how relationships substitute for local search and how they help bridge distance. Lastly, learning-by- hiring can be useful for extending the hiring firm’s geographic reach and access to remote knowledge (Song, Almeida, & Wu, 2003). In all these research designs, the existence and characteristics of networks affect the strength of the association between geography and knowledge. Connectivity as a Mediator of Geographical Effects on Knowledge A fourth stream of innovation research suggests that networks mediate the entire effect of geography on innovation. Empirically, patents are cited more frequently within the region in which they were invented than in others (Jaffe, Trajtenberg, & Henderson, 1993; Thompson & Fox-Kean, 2005). This finding supported the argu- ment that trajectories of technological knowledge are spatially sticky. However, new research designs were needed to examine how these local spillovers happened. Almeida and Kogut (1999) found that local spillovers did not occur equally across J. Glückler et al.